Evidence on the Incidence of Residential Property Taxes Across Households Evidence on the Incidence of Residential Property Taxes Across Households

Abstract - Property taxes are assessed by local taxing jurisdic- tions, including the county, city, and school district in which a property is located. This study examines the incidence of the resi- dential property tax burden across 357,264 owner–occupied homes in Dallas County, Texas. Progressivity indices are computed for total property taxes combined and for each jurisdictional tax (county, city, and school district). After allowing for the federal income tax deduction of property taxes, total property taxes combined are ap- proximately proportional. County and school taxes are proportional to slightly progressive, while city taxes are moderately regressive. Analysis suggests that tax rates contribute to the regressivity of city taxes because lower–income cities tend to have relatively higher tax rates. The homestead exemption affects the incidence of city and school taxes, but in opposite ways. It makes city taxes more regressive, but makes school taxes more progressive. The over–65 exemption increases progressivity for all three jurisdictional taxes.

INTRODUCTION

roperty taxes have always been the primary tax of local Pgovernments in the U.S. (Wallis, 2001). For fiscal year 1999, local governments in the U.S. collected total tax rev- enues of $316 billion, with property tax revenues represent- ing 72 percent ($228 billion) of that amount. Local govern- ments in the state of Texas collected total tax revenues of $23.6 billion in 1999, with 80 percent ($18.8 billion) of that amount being property tax revenues. Property taxation has long been relied on to fund local government needs, even though policymakers generally believe that property taxes are regres- sive (Wassmer, 1993; Musgrave, 2001). Despite this wide– spread belief, prior studies provide inconclusive evidence on the incidence of the property tax, with some studies finding property taxes to be regressive, and others finding them to Elizabeth Plummer be progressive (Suits, 1977; Phares, 1980; Metcalf, 1994). This paper uses property–specific data to provide evidence University of North on the incidence of residential property taxes assessed on Texas, Denton, TX homeowners, and to examine how institutional features (such 76203 as exemption amounts and rates) affect property tax inci- dence. Although the property tax’s design is similar across National Tax Journal local taxing jurisdictions, the institutional features differ. Vol. LVI, No. 4 Therefore, I examine the incidence of county, city, and school December 2003 district taxes separately. 739 NATIONAL TAX JOURNAL

The 2000 Certified Appraisal Master progressivity for all three jurisdictional File from the Dallas Central Appraisal taxes. This occurs because there is a dis- District provides a sample of 357,264 proportionate number of over–65 owner–occupied homes in Dallas County, homeowners in the bottom of the income Texas.1 The sample properties have a to- distribution, and because the exemption tal market value of approximately $44.7 amount decreases as property values in- billion, and paid a little over $850 million crease. in total property taxes for the 2000 tax By providing evidence on property tax year. The mean average property value is incidence and the distributional effects of $125,057, and the mean average property specific property tax features, this study tax payment is $2,381. The majority of this can help educate policymakers and other amount is for school taxes (about 59 per- parties interested in property tax reform. cent). County and city taxes comprise Property taxes are one of the most un- about 18 percent and 23 percent, respec- popular taxes (Dearborn, 1993; Baer, 1998), tively. and are frequently cited as being un- The results of this study are as follows. equally distributed across local jurisdic- After taking into account the deductibil- tions. While these perceived inequities ity of property taxes for federal income have led politicians and voters to press for tax purposes, results suggest that total a variety of property tax limitations, there property taxes for all jurisdictions com- remains little evidence and understand- bined are approximately proportional. ing of how these features combine to af- County and school taxes are proportional fect property tax incidence (Netzer, 2001). to slightly progressive, while city taxes are This study extends and improves upon moderately regressive. prior research on property tax incidence I examine the effect of three institutional in several ways. First, this study uses features on property tax incidence: tax property–level data to directly measure rates, the homestead exemption, and the property tax incidence. In general, other exemption for homeowners age 65 and studies use estimates of total U.S. prop- older. Tax rates have little–to–no effect on erty tax collections from the national in- the incidence of county and school taxes, come accounts data. Second, this study but do contribute to the regressivity of city examines property tax incidence sepa- taxes. Lower–income cities tend to have rately for counties, cities, and school dis- higher tax rates than middle– and high– tricts. Examining these taxes separately is income cities. The homestead exemption important because it enables me to deter- affects the incidence of city and school mine how tax rates, exemptions, and dif- taxes, but in opposite ways. It makes city ferences in income across different cities taxes more regressive because middle– and school districts affect the property tax income cities have smaller homestead ex- distribution. Prior studies lack the de- emption amounts than either low– or tailed information required for this level high–income cities. In contrast, the home- of analysis. Last, this study estimates the stead exemption makes school taxes more effect of federal tax deductibility on prop- progressive because the exemption erty tax incidence. amount (as a percentage of property The remainder of the paper is organized value) decreases as property values in- as follows. The next section reviews prior crease. The over–65 exemption increases research that examines property tax inci-

1 This study focuses only on persons who own their home and does not address the incidence of property taxes for renters (for example, see Carroll and Yinger (1994)). Property taxes are also levied on non–residential real estate, including commercial, industrial, and agricultural properties. I do not address the incidence of these taxes. 740 Evidence on the Incidence of Residential Property Taxes Across Households dence. Section 3 describes the data and to have school–aged children, may be light sample, including descriptive statistics. users of public services, and are more Section 4 presents the results, and Section likely to have smaller incomes.2 Therefore, 5 presents sensitivity analysis. Section 6 for each jurisdiction, a homeowner’s examines the features of the property tax property tax liability is calculated as fol- that affect its incidence. The final section lows:3 presents conclusions.

[1] PropTaxij = TaxValueij * tj PROPERTY TAXES = (AMVi – HSij – over65ij) * tj, Local governments in all 50 states levy where property taxes, and an additional 34 states levy a state–level property tax (Lorelli, PropTaxij = property tax liability of 2001). The property tax liability owed to residential property i to taxing a taxing jurisdiction is equal to the jurisdiction j, jurisdiction’s tax rate times the property’s TaxValueij = taxable value of residential taxable value. Taxable value is equal to the property i for taxing jurisdiction j’s property’s assessed market value (deter- purposes, mined by the local appraisal district), mi- AMVi = assessed market value of nus any allowable exemptions. It is com- residential property i, determined by mon for state governments to enact one the local appraisal district, set of property tax features that apply to HSij = amount of homestead exemption all local governments (for example, a stan- for residential property i allowed by dard exemption amount). In Texas, local taxing jurisdiction j, governments determine their own exemp- over65ij = amount of over–65 exemption tion amounts, subject to certain state– for residential property i allowed by imposed minimums. taxing jurisdiction j, and

The two most common exemptions are tj = jurisdiction j’s tax rate. the homestead exemption and the over– 65 exemption. The homestead exemption In some instances, a tax burden can be is only available for the residential prop- shifted from one taxpayer to another so erty where the homeowner resides. Nei- that the tax’s statutory incidence differs ther the homeowner nor spouse can claim from its economic incidence. Economists a homestead exemption on any other have posited three main theories about the property, including vacation homes or sec- economic incidence of property taxes: the ondary residences. The over–65 exemp- traditional view, the new (or capital tax) tion is available for homeowners aged 65 view, and the benefit view.4 In general, or older. These households are less likely these three theories make different predic-

2 There is also a disability homestead exemption available to persons who meet the definition of disabled under the Federal Old–Age, Survivor’s and Disability Insurance Act. If a person qualifies for both the over–65 and disability exemptions, s/he can only choose one. The 2000 County Report of Property Values (obtained from the Texas State Comptroller’s office) indicates that 88,697 households in Dallas County claimed the over–65 exemption, while only 9,535 claimed the disability exemption. 3 In addition to residential–related exemptions, Texas state law specifies that a homestead’s assessed market value (before exemptions) cannot be increased by more than 10 percent per year, times the number of years since the last appraisal. This 10 percent cap expires if the home is sold, and the homestead’s value can then be increased by more than 10 percent if necessary. To assess the impact of the 10 percent cap on tax incidence, I re–estimate the Suits indices assuming there is no legislative cap on annual appraisal increases. Results are essentially identical to those reported in the paper. 4 See Zodrow (2001) and Fullerton and Metcalf (2002) for a discussion and comparison of the three views of property tax incidence. 741 NATIONAL TAX JOURNAL tions about the taxpayer who ultimately Prior research provides conflicting con- bears the property tax burden. The tradi- clusions on property tax incidence, report- tional view argues that the property tax is ing Suits indices ranging from –0.13 to borne by both property owners and ten- 0.23. The differences in results likely arise ants. To eliminate possible shifts in the tax because of differences in the years exam- burden, I eliminate renter–occupied hous- ined, in the income measure used (annual ing and include only owner–occupied, resi- vs. lifetime), and in the method used to dential properties. The new view argues allocate taxes to a sample of households. that the property tax is borne primarily by Suits (1977) examines property tax inci- the owners of capital (Fullerton and dence for 1966 and 1970 using the total Metcalf, 2002). For purposes of this study, residential and commercial property tax I assume that residential property owners collections reported in the national income are the “owners of capital,” and that home accounts (Survey of Current Business, U.S. ownership proxies for capital ownership.5 Department of Commerce). The tax bur- The benefit view posits that the property den of approximately $25–30 billion is al- tax is not a tax but a user charge for local located across a stratified sample of 30,000 public services. Because the property tax households using the Survey of Economic is a voluntary price paid for goods and ser- Opportunity and an IRS file of 87,000 indi- vices received, the benefit view argues that vidual tax returns. Suits calculates an in- tax incidence is irrelevant (Fullerton and dex value of 0.23 for 1966, and 0.18 for Metcalf, 2002). The benefit view, however, 1970. These indices suggest that property depends on strong assumptions about zon- taxes are progressive, and that their ing and taxpayer mobility (Metcalf, 1994; progressivity decreased over the four– Zodrow, 2001), and its assumptions are year period examined. likely to be satisfied only in suburban ar- Phares (1980) also uses property tax eas, if at all (Fischel, 1995; Ladd, 1998, pp. data from the national income accounts. 34–5). In this study, I assume that tax inci- He allocates approximately $55 billion of dence is important and that the property residential and commercial property taxes tax burden falls on the property owner of across a stratified sample of about 200,000 owner–occupied housing. households using the Survey of Income and Studies examining property tax inci- Education, the Consumer Expenditure Sur- dence generally use the Suits index to vey, and an IRS file of 206,000 individual measure progressivity (Suits, 1977). The tax returns. He estimates Suits indices for Suits index estimates a tax system’s each of the 50 states for 1977, and finds progressivity by measuring the relation indices ranging from –0.23 to –0.06, with between the cumulative percentage of the an average value of –0.13. The Suits in- tax burden and the cumulative percent- dex for Texas is –0.06. In contrast to Suits age of total income. The index can vary (1977), Phares’ results suggest that the from –1 to +1, with –1 indicating maxi- property tax is regressive. mum regressivity and +1 indicating maxi- Metcalf (1994) examines the incidence mum progressivity. An index value of 0 of residential property taxes using the indicates a proportional tax. Consumer Expenditure Survey for 1984 and

5 Using home ownership as a measure of capital ownership seems to be a reasonable assumption for two rea- sons. First, I do not examine taxes imposed on non–residential (e.g., commercial and industrial) property, which are more likely to be shifted to owners of non–residential capital. Second, residential property values are correlated with capital income measures. For example, in the 1998 Texas data file, real estate taxes have a greater than 90 percent correlation with taxable interest income, tax–exempt interest income, dividends, and capital gains (all potential proxies for capital ownership). For the zip code data, residential property values have a greater than 70 percent correlation with taxable interest income. (The zip code data does not include details of the other income types.) 742 Evidence on the Incidence of Residential Property Taxes Across Households

1989. Of the 1,600 households surveyed, The final sample includes residential between 20 and 35 percent of the properties from 25 cities and 15 school homeowners in different regions of the districts. The largest city and school dis- U.S. reported no property tax liability trict are the city of Dallas and the Dallas (renters), and Metcalf (1994) must estimate Independent School District (ISD), with a property tax amount for each of these 157,804 and 137,459 properties, respec- households. To compute the Suits indices, tively. Metcalf (1994) uses two measures of in- Dallas County provides a 20 percent come: annual income and consumption. homestead exemption ($5,000 minimum), Using annual income, Metcalf (1994) finds a $69,000 over–65 exemption, and a that the property tax is regressive in both 0.505667 percent tax rate.7 If a taxing ju- 1984 and 1989 (Suits indices of –0.09 and risdiction offers a homestead exemption, –0.11, respectively). Using a consump- the State of Texas requires that the mini- tion–based definition of income, however, mum amount be $5,000. Of the 25 cities, his evidence suggests that property taxes 15 do not allow a homestead exemption. are less regressive in 1984 (Suits index of Seven of the cities (including Dallas) al- –0.04) and slightly progressive in 1989 low a 20 percent homestead exemption. (Suits index of 0.01). One city allows a 10 percent homestead exemption, and two cities allow a 1 per- DATA AND SAMPLE cent homestead exemption. All of the cit- ies provide for an over–65 exemption. The The 2000 Certified Appraisal Master amounts range from a low of $3,000 to a File, purchased from the Dallas Central high of $75,000, and the median amount Appraisal District (DCAD), contains in- is $30,000. (The over–65 exemption is formation for 623,032 residential proper- $64,000 for the city of Dallas.) The prop- ties in Dallas County. DCAD is respon- erty tax rates also vary widely across cit- sible for appraising all properties in Dal- ies, with a range from 0.2462 to 0.7900 las County. From the total file, I delete percent. The median tax rate is 0.6675 per- properties that are not classified as single– cent. family residences (114,802 observations), The state of Texas mandates that all properties that are under construction or school districts must grant a $15,000 in a special taxing district (12,301 obser- homestead exemption. Five of the 15 vations), properties that are in cities with school districts in the final sample (includ- less than 100 properties or are not classi- ing Dallas ISD) also offer a percentage– fied as within a city or school district based homestead exemption that a tax- (19,567 observations), and properties with payer receives in addition to the $15,000 no homestead exemption (119,098 obser- mandatory exemption amount. The per- vations).6 This results in a final sample of centages range from 10 to 20 percent of a 357,264 owner–occupied properties. Each property’s market value. All school dis- property is a single–family residence that tricts provide for an over–65 exemption, is classified as a homestead, and is taxed with amounts ranging from $10,000 to by three separate taxing jurisdictions: $45,000. The median over–65 exemption Dallas County, and the city and school is $10,000, and Dallas ISD’s over–65 ex- district in which the residence is located. emption is the largest ($45,000). School tax

6 Homestead exemptions can only be claimed for the residential property where the homeowner resides. I delete properties with no homestead exemption because I cannot assume that the homeowner occupies the property (i.e., the homeowner may be using it as rental property or as a vacation home). 7 The total county tax rate is the sum of tax rates from the following jurisdictions: Dallas County (0.196), School Equalization (0.005667), Parkland Hospital (0.254), and Dallas County Community College (0.05). 743 NATIONAL TAX JOURNAL rates are about three times the amount of tricts. The remaining property taxes were the county and city tax rates, and do not paid to cities ($198 million) and to Dallas vary much across school districts (from County ($151 million). 1.2467 to 1.6950 percent). The median rate Table 2 presents descriptive statistics for is 1.548 percent. the sample properties individually. The Table 1, Panel A, provides summary first row shows that the mean assessed information on the sample’s total property market value for all properties is $125,057, values relative to all residential properties while the median value is $87,620. Al- in Dallas County. The 357,264 sample though the range of property values is properties have a total market value of wide, one–half of the sample homes have almost $44.7 billion. Relative to all resi- an assessed value between $61,440 and dential properties in Dallas County, the $135,160. The next three rows of Table 2 sample represents 61.3 percent of all prop- provide information on the taxable values erties and 75.3 percent of the total market for county, city, and school district pur- value. The sample’s taxable values range poses. The median county and school dis- from $29.8 billion to $33.6 billion. The tax- trict taxable values are both about $62,000, able values for city and school district pur- while the median city taxable value is poses are slightly larger than for county $71,448. The larger taxable value for cities purposes because of the smaller exemp- arises because the majority of cities (15 of tions that cities and school districts pro- 25) do not allow a homestead exemption. vide. Panel B provides information on the In contrast, the county and all school dis- total property taxes paid by the sample tricts provide a homestead exemption. The properties for the year 2000, and the clas- minimum taxable value for all three juris- sification of those taxes as paid to the dictions is zero, reflecting the fact that the county, city, or school district. In total, the over–65 exemption can reduce a property’s sample properties paid a little over $850 taxable value to no less than zero. million in property taxes, with $501 mil- The last four rows of Table 2 describe lion (59.0 percent) going to the school dis- property taxes paid. On average, home-

TABLE 1 SUMMARY INFORMATION ON TOTAL PROPERTY VALUES AND PROPERTY TAXES PAID Panel A: Total Property Values of Sample As Reported by Sample as a Percentage Sample Dallas County1 of Dallas County Number of Residential Properties 357,264 582,394 61.3%

Market Value $44,678,424,130 $59,340,997,920 75.3%

Taxable Values:

County $29,797,148,307 $43,742,651,998 68.1% City 33,599,311,888 51,365,353,873 65.4 School District 31,843,223,054 47,172,441,676 67.5 Panel B: Total Property Taxes Paid by Sample As a Percentage of Amount Total Property Taxes County Taxes $ 151,141,679 17.7% City Taxes 197,910,080 23.3 School Taxes 501,548,640 59.0

Total Property Taxes $ 850,600,399 100.0% 1 Data for Dallas County is taken from the Estimated Value Report for Tax Year 2000 (://www.dallascad.org/ 2000certified.htm).

744 Evidence on the Incidence of Residential Property Taxes Across Households

TABLE 2 DESCRIPTIVE STATISTICS FOR INDIVIDUAL PROPERTIES (N = 357,264) Variable Mean Median StDev Min Q1 Q3 Max Market Value $125,057 $87,620 $164,076 $2,700 $61,440 $135,160 $16,360,790

Taxable Value: County $83,404 $61,777 $123,161 $0 $31,288 $97,176 $8,783,200 City 94,046 71,448 123,893 0 39,136 111,704 8,783,200 School District 89,131 62,770 129,002 0 34,320 102,893 9,866,100

Property Taxes: Total $2,381 $1,723 $3,260 $0 $942 $2,777 $255,722 County 423 313 623 0 160 492 44,414 City 554 434 659 0 244 673 58,628 School District 1,404 982 2,026 0 537 1,629 152,681 owners paid property taxes totaling $2,381 Measuring Household Income (mean) or $1,723 (median) for the 2000 tax year. One–half of the homeowners paid The first step in calculating S is to rank between $942 and $2,777. On average, the sample households according to in- homeowners paid $1,404 for school taxes, come. Lifetime income measures are pref- $423 for county taxes, and $554 for city erable to annual income measures because taxes. lifetime measures are less subject to tem- porary fluctuations (Metcalf, 1994, pp. 65– 6), and because individuals are likely to RESULTS make consumption decisions based on I use the Suits index to measure prop- lifetime income and not annual income erty tax progressivity. The Suits index (S) (Fullerton and Metcalf, 2002). Consistent is constructed from a graph of the cumu- with prior research (Poterba, 1989, 1991; lative tax burden (y axis) against cumula- CBO, 1990; Metcalf, 1994), I use consump- tive income (x axis): tion as a proxy for lifetime income, and use residential property value to measure 10 consumption. To calculate S, I rank house- Σ —1 (PT + PT )(Y –Y) 2 k k–1 k k–1 holds according to residential property [2] S = 1 – ————————————k=1 value, divide them into deciles, and use 5,000 cumulative property value as the measure of Y in equation [2] above. where I use a property’s assessed market value

PTk = cumulative percent of property (available from the DCAD file) to proxy tax burden for population deciles 1 for its market value.8 Because S is based through k, and on property value rankings rather than

Yk = cumulative percent of total income absolute values, differences between a for population deciles 1 through k. property’s assessed market value and its actual market value should not bias S un- Negative index values indicate a regres- less assessed–to–actual market value var- sive tax, 0 indicates a proportional tax, and ies systematically with value. The main positive values indicate a progressive tax. reason for differences between assessed

8 Assessed market value used for ranking purposes is the DCAD market value without regard to the legislative cap, if applicable (see footnote 3 above).

745 NATIONAL TAX JOURNAL value and market value is infrequent as- erty tax burden after deducting the ex- sessment (O’Sullivan, Sexton, and pected federal income tax benefit due to Sheffrin, 1995, p. 37). If property values itemizing). This expected net property tax are increasing over time (as in Dallas payment (NetPropTax) is calculated as: County), then properties with less fre- quent assessments will generally have [3] NetPropTaxi = PropTaxij * pr(item)i * lower assessment ratios. To examine (1 – MTR ) + PropTax (1 – pr(item) ), whether the frequency of assessment var- i ij * i ies systematically across property values where for this study’s sample, I compute the av- erage appraisal lag across property value PropTaxij = actual property taxes paid by deciles. Appraisal lag is defined as the household i to taxing jurisdiction j, most current appraisal year minus the pr(item)i = probability that household prior appraisal year. i itemizes property taxes on its Results suggest that lower–value prop- federal income tax return, and erties are assessed less frequently than MTRi = household i’s marginal tax rate. higher–value properties (the correlation coefficient between appraisal lag and prop- I then use NetPropTaxi as the measure of erty value is –0.229, and the mean appraisal PT in equation [2] above. lag monotonically decreases across the ten The probability that a household item- property value deciles).9 If infrequent as- izes depends upon the household’s in- sessments lead to lower assessment ratios, come level. Because information about then the DCAD assessment practices may income is not publicly available for indi- cause assessment ratios to be lower for vidual households, I estimate income for low–value properties than for high–value each household using the Internal Rev- properties. Accordingly, the Suits indices enue Service 1998 Zip Code Data File (the may understate property tax progressivity most current year available). This file pro- somewhat. In the sensitivity analysis sec- vides aggregate AGI and number of re- tion, I examine the effects on the Suits in- turns, by zip code. There are 212 zip codes dex of this possible measurement error in in the sample, with an average of 1,685 appraisal values. properties in a zip code. For each zip code, I estimate the average AGI per household Measuring the Property Tax Burden (aggregate AGI divided by number of re- turns). For the sample households, the The second step in computing S is to AGI estimates range from $21,668 to determine the property tax burden for $671,866, with a median value of $49,825. each household. Property taxes are de- Both the mean and median AGI values ductible for federal income tax purposes, monotonically increase across property thereby decreasing the net tax burden of value deciles, and the Spearman correla- those taxpayers who can itemize. To al- tion coefficient between property value low for the effect of the federal tax deduc- and AGI is 0.70. tion on property tax incidence, I compute Each household’s AGI estimate is then a measure of the expected net property tax matched with its corresponding AGI burden for each household (i.e., the prop- bracket from the Internal Revenue Service

9 The mean (median) appraisal lag is 7.8 (8) years for the lowest decile and 3.5 (2) years for the highest decile. Appraisal lags are significantly different across deciles (α < 0.001 using the Kruskal–Wallis test). The Kruskal– Wallis test is a nonparametric test that tests whether the distribution of a variable is the same across the different deciles. No assumptions about the actual form of the probability distributions are required (Hol- lander and Wolfe, 1973). 746 Evidence on the Incidence of Residential Property Taxes Across Households

1998 Texas Data File. For each of 12 dif- explore the effects of measurement error ferent AGI brackets, this file provides (1) in a later section. the average income tax liability, and (2) the percentage of taxpayers that report Results real estate taxes on their Form 1040, Sched- ule A. Information from the correspond- Table 3, Panel A, presents the median ing AGI bracket is used to calculate MTRi property value, the percentage of property and pr(item)i for each household. For the value owned, and the percentage of net sample, pr(item)i ranges from 0.09 to 75.1 property taxes paid by each decile. House- percent. These values are used in equa- holds in the lowest decile have a median tion [3] to provide an estimate of the prop- property value of $30,420, while those in erty tax burden for each household, after the highest decile have a median property deducting the expected federal income tax value of $322,095. Panel A, column 3, benefit due to itemizing. It should be shows that the wealthiest 20 percent of the noted that this method of estimating households (deciles 9 and 10) own 49.3 household income fails to capture varia- percent of total property value, and paid tions in the probability of itemizing and 49 percent of the total net property tax bur- in marginal tax rates that exist across den. (They paid 49.9, 45.1, and 50.3 per- households within a given zip code. For cent of county, city, and school taxes, re- example, to the extent a high–income spectively.) Panel B of Table 3 shows the household lives in a low–income neigh- cumulative percentages of property val- borhood, the estimate of pr(item) and MTR ues and net property tax payments by for that household will be too low. I decile. For example, the poorest 20 per-

TABLE 3 PROPERTY VALUE, PERCENTAGE OF PROPERTY VALUE OWNED, AND PERCENTAGE OF NET TAXES PAID (BY DECILE), AND SUITS INDICES1 Total Sample Median Property Property Property County City School Decile Value Value Taxes Taxes Taxes Taxes Panel A: Absolute Percentages 1 $30,420 2.3% 1.3% 1.8% 1.9% 0.9% 2 49,100 3.9 3.2 3.5 3.9 2.8 3 61,440 5.0 4.4 4.6 5.0 4.2 4 72,450 5.7 5.8 5.6 6.3 5.6 5 82,500 6.6 6.9 6.6 7.5 6.8 6 93,360 7.5 8.1 7.6 8.6 8.0 7 109,730 8.8 9.6 9.1 10.0 9.6 8 135,160 10.9 11.7 11.3 11.7 11.8 9 179,210 14.6 15.5 15.1 15.5 15.6 10 322,095 34.7 33.5 34.8 29.6 34.7 Panel B: Cumulative Percentages 1 $30,420 2.3% 1.3% 1.8% 1.9% 0.9% 2 49,100 6.2 4.5 5.3 5.8 3.7 3 61,440 11.2 8.9 9.9 10.8 7.9 4 72,450 16.9 14.7 15.5 17.1 13.5 5 82,500 23.5 21.6 22.1 24.6 20.3 6 93,360 31.0 29.7 29.7 33.2 28.3 7 109,730 39.8 39.3 38.8 43.2 37.9 8 135,160 50.7 51.0 50.1 54.9 49.7 9 179,210 65.3 66.5 65.2 70.4 65.3 10 322,095 100.0 100.0 100.0 100.0 100.0 Suits Index: 0.006 0.012 –0.047 0.026 1 This table uses net property tax payments (NetPropTaxi) as the measure of PT in equation [2]. NetPropTaxi is a household’s estimated property tax burden after deducting the expected federal income tax benefit due to item- izing. 747 NATIONAL TAX JOURNAL cent of households own 6.2 percent of to- then the estimated net tax burden will be tal property value, but paid only 4.5 per- too low. The effect of measurement error cent of total net property taxes. (They paid on the Suits index depends on the direc- 5.3, 5.8, and 3.7 percent of county, city, and tion and magnitude of the error, as well school taxes, respectively.) as whether the error differs across the in- The last line of Table 3 presents Suits in- come distribution. To test for the effects dices for total property taxes combined and of possible measurement error in AGI, I for each of the jurisdictional taxes. The adjust AGI in several ways and estimate Suits index for total property tax payments the Suits indices for each scenario. is 0.006, suggesting that property taxes are The most likely source of measurement approximately proportional. For the sepa- error in AGI arises because my estimation rate jurisdictions, the Suits indices for method does not capture income varia- county and school taxes are 0.012 and tions within neighborhoods. There are 0.026, respectively. This suggests that likely to be high–income households in county and school taxes are proportional low–income neighborhoods, as well as to slightly progressive. In contrast, the Suits low–income households in high–income index for city taxes is –0.047, suggesting neighborhoods. For these households, that city taxes are moderately regressive. AGI will be measured with error. If the The regressivity of city taxes arises prima- measurement error in AGI is not system- rily because the top decile owns 34.7 per- atically different across property value cent of the property value, but has a city deciles, but only adds noise to the net tax burden of only 29.6 percent. In contrast, property tax measure in equation [3], the households in each of deciles 4 through 9 Suits index values are unaffected. I run have a net tax burden greater than their several simulations by adding noise to the proportionate share of property value. For data, and the Suits indices are essentially example, households in decile 7 own 8.8 identical to those in Table 3. percent of the property value, but have a A more serious problem for drawing net city tax burden of 10.0 percent.10 inferences could occur if AGI measure- ment error is systematically different for SENSITIVITY ANALYSIS households in the upper versus lower ends of the income distribution. For ex- As discussed above, I must estimate ample, if AGI is underestimated more (or household AGI in order to determine the overestimated less) for high–income probability that a household itemizes households than for low–income house- property taxes on its federal tax return. holds, then the Suits indices will be biased Using an AGI estimate could potentially upward. Conversely, if AGI is underesti- affect the Suits indices calculations. A mated more (or overestimated less) for household’s probability of itemizing gen- low–income households, the Suits indices erally increases as AGI increases. There- will be too low. fore, if a household’s AGI estimate is too To examine the bias from measurement low, then the probability of itemizing will error that differs systematically across be too low, and the estimated net prop- deciles, I run simulations by adjusting erty tax burden will be too high. Con- AGI upward for households in the top five versely, if the AGI estimate is too high, deciles of Table 3, and then for households

10 I also rank households using AGI as the measure of income, divide the households into deciles, and compute the Suits indices using cumulative AGI as the measure of Y in equation [2]. Results are qualitatively similar to those reported in Table 3, with values of 0.005, –0.044, and 0.012 for county, city, and school taxes, respectively. The Suits indices computed in this manner are not affected by possible systematic differences in assessment ratios. 748 Evidence on the Incidence of Residential Property Taxes Across Households in the lower five deciles.11 As expected, the with an average increase of about 0.024. first adjustment results in less progressive (Recall that appraisal lag was more severe Suits indices than those in Table 3, and the for low–value properties.) The relative dif- second adjustment results in more pro- ferences between the jurisdictional indi- gressive indices. However, in both cases, ces remain approximately the same. the magnitude of the bias is relatively small. The largest effects change the Suits FEATURES OF THE PROPERTY TAX indices by about 0.016 to 0.019. In addi- THAT AFFECT ITS INCIDENCE tion, because the changes in Suits values are similar across jurisdictions, the rela- Local Features tive differences between the jurisdictional There are three institutional features indices remain approximately the same.12 that can affect property tax incidence: tax I also examine the effects of possible rates, the homestead exemption, and the measurement error in appraisal values. As over–65 exemption. Table 4 provides in- discussed earlier, the main reason for dif- formation on how these features differ ferences between assessed value and mar- across decile groups for each of the juris- ket value is infrequent assessment dictional taxes. Panel A shows that the (O’Sullivan et al., 1995, p. 37). Given the average county tax rate is identical across increasing property values in Dallas decile groups (because all households in County during this time period, proper- Dallas county are taxed at the same county ties with less frequent assessments will tax rate), and that the average school tax have assessed market values that are gen- rates are similar across all decile groups. erally too low. Using property value as a Accordingly, differences in tax rates have proxy for lifetime income will therefore little–to–no effect on the tax incidence of understate income for these households. county and school taxes. In contrast, Panel To examine the effects of measurement A shows that the average city tax rate is error in appraisal values, I re–estimate the highest for the lowest–valued properties, Suits indices after adjusting property val- indicating that tax rates contribute to the ues for appraisal lags. Specifically, I ad- regressivity of city taxes. The average city just property values upward by 10 per- tax rate is 0.6428 percent for deciles 1 cent times the number of years since the through 3, versus 0.6072 and 0.6034 per- last appraisal, and then re–estimate the cent for the other decile groups. Suits indices using these adjusted prop- Panel B presents information on the erty values to rank households. As ex- average homestead exemption amount as pected, the recomputed Suits indices are a percentage of property value (i.e., the slightly more progressive than in Table 3, percentage of the property’s market value

11 Specifically, I double AGI for the top 10 percent of households, as well as the top 20, 30, 40, and 50 percent, and do the same for the lowest 10, 20, through 50 percent of households. For the upper–income adjustments, the largest difference comes from doubling AGI for the top 20 percent of households. This results in Suits values of –0.003, –0.065, and 0.011, for county, city, and school taxes, respectively. For the lower–income adjustments, the largest difference comes from doubling AGI for the bottom 50 percent of households, resulting in Suits values of 0.031, –0.027, and 0.044, for county, city, and school taxes, respectively. 12 My analysis does not allow for the probability of itemizing to vary across taxpayer age groups, and there is some evidence that taxpayers aged 65 and older are less likely to itemize. For 1999, 32.0 percent of taxpayers under age 65 itemized, while only 29.4 percent of taxpayers age 65 and older itemized. I re–run all my analy- ses assuming that no over–65 taxpayers itemize, and also assuming that only the wealthiest taxpayers over– 65 itemize (i.e., those in the upper 10 or 20 percent of households). All results are essentially identical to those reported in the paper. These results do not differ appreciably because the probability that over–65 taxpayers itemize is already fairly low. Almost 40 percent of over–65 taxpayers are in deciles 1 through 3, and the me- dian probability of itemizing for each of these deciles is between 6.6 and 17.2 percent. I thank Sandra Byberg at the IRS Statistics of Income Office for providing the 1999 information. 749 NATIONAL TAX JOURNAL

TABLE 4 FEATURES OF THE PROPERTY TAX THAT AFFECT ITS INCIDENCE County Taxes City Taxes School Taxes Panel A: Statutory Tax Rates Deciles 1 through 3 0.5057%1 0.6428%2 1.5578%1 Deciles 4 through 7 0.5057 0.6072 1.5728 Deciles 8 through 10 0.5057 0.6034 1.5847

Full Sample 0.5057% 0.6168% 1.5719% Panel B: Homestead Exemption as a Percentage of Market Value Deciles 1 through 3 21.0%1 14.2%2 44.1%2 Deciles 4 through 7 20.0 8.5 23.1 Deciles 8 through 10 20.0 13.9 18.4

Full Sample 20.3% 11.8% 28.0% Panel C: Over–65 Exemption as a Percentage of Market Value3 Deciles 1 through 3 78.5%2 77.0%2 38.9%2 Deciles 4 through 7 74.4 56.8 27.8 Deciles 8 through 10 36.6 30.0 14.9

Full Sample 66.0% 57.5% 28.7% Panel D: Suits Indices Not Allowing for the Effects of Federal Income Tax Deductibility4 Full Sample 0.051 –0.012 0.062 1Differences across decile groups are not significant at α < 0.001 using the Kruskal–Wallis test (Hollander and Wolfe, 1973). 2Differences across decile groups are significant at α < 0.001 using the Kruskal–Wallis test. 3Panel C only includes properties owned by taxpayers age 65 or older. 4The Suits Indices are computed using actual (or nominal) tax payments as the measure of PT in equation [2]. that is exempt from tax). The homestead In contrast to city taxes, the homestead exemption contributes little to making exemption increases the progressivity of county taxes more progressive because all school taxes. The last column of Panel B properties in Dallas County are allowed shows that the homestead exemption de- a 20 percent homestead exemption, with creases as property values increase, a $5,000 minimum. For both cities and averaging about 44 percent of market school districts, however, homestead ex- value for properties in the bottom deciles, emptions differ significantly across decile 23.1 percent for the middle deciles, and groups. The homestead exemption con- 18.4 percent for the top deciles. As man- tributes to the regressivity of city taxes dated by the state of Texas, all 15 school because middle–decile households re- districts in the sample provide for a ceive a smaller percentage homestead ex- $15,000 homestead exemption. A fixed emption than upper–decile households. exemption amount increases school tax The middle column of Panel B shows that progressivity because it decreases the ef- the homestead exemption averages about fective tax rate for lower–valued proper- 14 percent of market value for properties ties more than it does for higher–valued in the bottom three and top three deciles, properties.13 and only about 8 percent of property value The over–65 exemption increases tax for the middle deciles. Fifteen of the 25 progressivity for all three jurisdictional cities represented by the sample provide taxes for two reasons. First, there is a dis- no homestead exemption, and a large proportionate number of over–65 number of these households fall in the homeowners in the bottom part of the dis- middle of the income distribution. tribution. For the lowest three deciles,

13 Five of the school districts also allow taxpayers to take a percentage–based homestead exemption. 750 Evidence on the Incidence of Residential Property Taxes Across Households about 34 percent of homeowners are over emption amounts range from $1,000 to full age 65, while for the middle and upper exemption (Baer, 1998). This data does not deciles, only about 22 percent of reflect exemptions that are granted at the homeowners are over age 65. Second, all option of local governments. Therefore, jurisdictions have fixed over–65 exemp- the use of general and special homestead tion amounts, and this disproportionately exemptions appears prevalent. Results of benefits lower–income homeowners. this paper suggest that exemptions are ef- Panel C of Table 4 provides information fective at increasing property tax on the average over–65 exemption as a progressivity when they are a fixed percentage of property value. For all ju- amount, but that differences in exemption risdictions, the over–65 exemption (as a amounts across municipalities can con- percentage of property value) decreases tribute to tax regressivity. Exemptions for as property values increase, and this pat- the elderly also increase progressivity be- tern is most pronounced for county and cause older homeowners are concentrated city taxes. For the poorest over–65 in the bottom of the income distribution. homeowners, the average county and city For exemptions to be effective, however, exemption amount exceeds 75 percent of eligible homeowners must be aware of property value, while for the wealthiest and obtain the exemption. Baer (1998) ex- over–65 homeowners, the county and city amines elderly taxpayers in six states and exemptions average about 36 and 30 per- estimates that participation rates in ex- cent of property value, respectively. For emption programs range from 60 to nearly school districts, the exemption averages 100 percent. Baer (1998) also finds that par- about 39 percent of property value for the ticipation rates are lower for low–income poorest over–65 households, and only households. Although a program may be about 15 percent for the wealthiest. These well–designed, poor administration could results reflect the relative generosity of the reduce the exemption’s effectiveness at over–65 exemption amounts offered by increasing tax progressivity. each jurisdiction. Dallas County’s over– 65 exemption is a fixed amount of $69,000. Effects of Federal Tax Deductibility All cities offer an over–65 exemption, with amounts ranging from $5,000 to $75,000. The deductibility of property taxes for The median exemption amount for the income tax purposes causes them to be sample properties is $50,000. Most school less progressive because the beneficiaries districts offer an over–65 exemption of of the reduced liability are generally con- only $10,000, although Dallas ISD (repre- centrated in the upper tiers of the income senting 38.5 percent of the sample) offers distribution. These property owners are a $45,000 exemption amount. more likely to itemize, pay larger prop- All states (except Kansas and Missouri) erty tax amounts, and have higher mar- and the District of Columbia mandate ginal tax rates. For the sample households, some form of a homestead exemption the average probability of itemizing in- and/or an over–65 exemption (Duncombe creases monotonically across deciles, from and Yinger 2001). Forty–five states offer 6.6 percent for decile 1 to 70.5 percent for special exemptions that apply to elderly decile 10. To measure the effects of fed- homeowners as well as other possible eral tax deductibility on property tax in- groups (e.g., the disabled and/or veter- cidence, I re–compute the Suits indices ans), and twenty states offer a general using actual (or nominal) tax payments as homestead exemption to all homeowners. the measure of PT in equation [2]. Results Sixteen states offer both general and spe- are presented in Panel D of Table 4. Not cial exemptions. The state–mandated ex- allowing for federal tax deductibility in- 751 NATIONAL TAX JOURNAL creases all Suits indices by a little less than Acknowledgments 0.04 points. For total property taxes com- bined, the recomputed Suits index is 0.043 I thank Karen A. Smith for helpful com- compared with 0.006 with deductibility ments on an earlier version of this paper, (see Table 3). and appreciate the very detailed com- ments and suggestions from two anony- mous reviewers and the editor, Therese CONCLUSIONS McGuire. I am especially indebted to By examining residential property taxes David Holmes of the University of Texas for 357,264 owner–occupied homes in at Dallas, Information Resources, for his Dallas County, Texas, this study provides unselfish work in helping with data ex- evidence on the incidence of residential traction from the DCAD tapes. property taxes, and on how institutional features affect that incidence. After allow- REFERENCES ing for the federal income tax deduction of property taxes, total property taxes Baer, David. combined are approximately propor- “Awareness and Popularity of Property Tax tional. County and school taxes are pro- Relief Programs.” Public Policy Institute portional to slightly progressive, while Working Paper. Washington, D.C.: Ameri- city taxes are moderately regressive. Tax can Association of Retired Persons, 1998. rates contribute to the regressivity of city Carroll, Robert J., and John Yinger. taxes because lower–income cities tend to “Is the Property Tax a Benefit Tax? The Case have relatively high tax rates. The home- of Rental Housing.” National Tax Journal 47 stead exemption makes city taxes more No. 2 (June, 1994): 295–316. regressive, but makes school taxes more Congressional Budget Office (CBO). progressive. The over–65 exemption in- Federal Taxation of Tobacco, Alcoholic Beverages, creases progressivity for all three jurisdic- and Motor Fuels. Washington, DC: Govern- tional taxes. ment Printing Office, 1990. The property tax remains one of the Dearborn, P. most unpopular taxes because of per- “ACIR 1993 Poll Takes Pulse on Taxes.” State ceived inequities. This study provides re- Tax Notes 5 (1993): 781. searchers and policymakers with evidence Duncombe, William, and John Yinger. on the distribution of the property tax “Alternative Paths to Property Tax Relief.” In burden, and also on the distributional ef- Property Taxation and Local Government Finance, fects of specific property tax features. This edited by Wallace E. Oates, 243–300. Cam- evidence can educate those parties who bridge, MA: Lincoln Institute of Land Policy, are interested in property tax reform about 2001. the ultimate impact of a proposed change Fischel, William A. to the property tax’s design. A useful ex- Regulatory Takings: Law, Economics, and Poli- tension of this study would be to make tics. Cambridge, MA: Harvard University hypothetical changes to different property Press, 1995. tax features and estimate the redistribu- Fullerton, Don, and Gilbert E. Metcalf. tion that would occur if such proposals “Tax Incidence.” In Handbook of Public Eco- were enacted. One could also examine al- nomics, Vol. 4, edited by Alan Auerbach and ternatives to the property tax (i.e., local Martin Feldstein. Amsterdam: Elsevier, 2002. sales and income taxes), and compare the Hollander, Myles, and Douglas Wolfe. incidence of these alternative taxes with Nonparametric Statistical Methods. New York: that of the property tax. John Wiley and Sons, 1973.

752 Evidence on the Incidence of Residential Property Taxes Across Households

Ladd, Helen F. Phares, Donald. Local Government Tax and Land Use Policies in Who Pays State and Local Taxes? Cambridge, the United States. Northampton, MA: Ed- MA: Oelgeschlager, Gunn, and Hain, ward Elgar Publishing, 1998. 1980. Lorelli, Michael F. Poterba, James. Special Report: State and Local Property Taxes “Lifetime Incidence and the Distributional (August 2001, No. 106). Washington, D.C.: Burden of Excise Taxes.” American Economic Tax Foundation. Review 79 No. 2 (May, 1989): 325–30. Metcalf, Gilbert E. Poterba, James. “The Lifetime Incidence of State and Local “Is the Gasoline Tax Regressive?” In Tax Taxes: Measuring Changes During the Policy and the Economy: Volume 5, edited by 1980s.” In Tax Progressivity and Income In- David F. Bradford, 145–64. Cambridge, MA: equality, edited by Joel Slemrod, 59–94. Cam- The MIT Press, 1991. bridge, England: Cambridge University Suits, Daniel. Press, 1994. “Measurement of Tax Progressivity.” Ameri- Musgrave, Richard A. can Economic Review 67 No. 4 (September, “Commentary on Local Property Taxation 1977): 747–52. in Theory and Practice.” In Property Taxa- Wallis, John J. tion and Local Government Finance, edited by “A History of the Property Tax in America.” Wallace E. Oates, 339–42. Cambridge, MA: In Property Taxation and Local Government Lincoln Institute of Land Policy, 2001. Finance, edited by Wallace E. Oates, 123–47. Netzer, Dick. Cambridge, MA: Lincoln Institute of Land “Local Property Taxation in Theory and Policy, 2001. Practice.” In Property Taxation and Local Gov- Wassmer, Robert W. ernment Finance, edited by Wallace E. Oates, “Property Taxation, Property Base, and 321–38. Cambridge, MA: Lincoln Institute Property Value: An Empirical Test of the of Land Policy, 2001. ‘New View.’” National Tax Journal 46 No. 2 O’Sullivan, Arthur, Terri A. Sexton, (June, 1993): 135–60. and Steven M. Sheffrin. Zodrow, George R. Property Taxes and Tax Revolts: The Legacy of “The Property Tax as a Capital Tax: A Room Proposition 13. Cambridge, England: Cam- with Three Views.” National Tax Journal 54 bridge University Press, 1995. No. 1 (March, 2001): 139–56.

753